PHR-803T batch testing.. Windowless 803Ts?

I'm just going to increase my diodes until they either die or output at least 200mW. Look at this, the third diode I pushed *drool*:



The second line is my running efficiency difference curve, used for detecting knees.

did you end up killing that 300 mw + unit ???

I had some funny diodes in the last batch too...

A couple had mode issues. They had an absurd efficiency at low currents, i classified them as freaks, set to high current, and the power came out low, and the beam deformed. I lowered the current, to find the one, where it still made a nice beam (118mW) and the power was the same as at 160mA - 125mW!

Some which i clasified as weak, i set to a medium high current of 143mA, and they climbed in power for a while, untill they stabilised at a power i would expect from a freak.

Just when i think i see a pattern, the PHRs pull another trick, and do something unexpected.


But i noticed an interesting pattern. Originally i assumed, the freaks would be able to live longer at the same power as lower efficiency diodes, due to taking less current, heating up less and because the high efficiency could indicate a more perfect die. But every single freak dies when pushed!

And the weak ones, the ones i used to think of as useless are turning out to be good.. These are weird diodes.

Looks like Dave is not the only one with a 300mW blu ray anymore...  ;D

Man, i have some diodes here (same GB as yours), i need to make me a 300mW blu ray! One of them is bound not to blow on powerup!

But first i'll start with the ones i deemed too weak to use from the old batch!



The massacre begins!

Gallium nitride is some crazy stuff, and the processes to make diodes are also quite crazy. With everything still so relatively new, I'm not surprised at all at how much they vary. And remember, these are JUST the ones that passed QC and the whole testing process. Imagine the crazy things that happened with all the OTHER diodes.

Gallium nitride really is amazing stuff by itself, even the fact that it works at all is still a little baffling. Gallium nitride has a relatively HUGE defect density, even when done perfectly, but yet it still works better than it ever really should. For instance, silicon wafers that computer chips are made on have a defect density of less than 1/sq. cm. Gallium nitride can still have a defect density greater than 10^10 /sq. cm, and it still works. Amazing stuff.

Even now, people really have relatively little understanding at all of this stuff, or the other III-nitrides. InN is a funny example: it took decades for people to realize that the bandgap that has been "known" was completely wrong, by a factor of almost 3. Even something as basic as a bandgap, and there are still people fighting over what the bandgap really is for InN.

When a few atoms make all the difference in the world, it's hard to get repeatability. Here, the results of what comes out of the reactor can change dramatically on a day-to-day basis. And since it changes day-to-day, but you won't have any results on anything for weeks or months later, it's a tough thing to trace problems to where they actually happened. When something is weird, was it the growth? Was it the design of the device? Was it a processing step? Is it the contacts? Facets? Huge numbers of variables.

Wow!

I had a distorted beam diode (mode issues) that had a completelly deformed beam at 166mA.

I just set it to 251mA and the spot is suddenly round again! Before, the beam was like a razor!!
It's actually doing better at 251mA then it was at 166mA!


And the power! 230mW! OMG!


I'm starting to think, that my dead freaks died from being too pampered.. ;D

^This is making me happy.

It probably has those same great kinks and changes in slope, as well. If you can, compare the slope of the P/I chart in each section between its kinks, see if it's more efficient in one or the other.

Btw, this diode was only doing 130mW at 166mA, altho the simplified efficiency test showed it to be a high efficiency diode..

Now it's making my 6x look dim! I think the wavelength shifted higher. At least the filter test would suggest so.. Is that possible? And i don't mean from heat. It's in a build, that can keep an open can cool for at least an hour, if not more!



EDIT: It just survived >20 minutes of my usual (ab)use... No duty cycle, left on for as long as i like, powered up and down several times, often for just a short period, almost a blink, which i always feel like it's the worst thing to do..

Basically, i just used it, without thinking how high the current is and how careful i should be..

I just measured it again, 231mW, 231mW, just like before.. No degradation noticable... It may just outlive my previous one, that was in this build at 166mA, but only did 159mW..

Man, i never had to set the range on my meter to high for a blu ray before!  ;D


I thought its a bad diode, that needs to be set lower, because of the mode issue, not higher! I just wanted to kill it to make room for another....  :
I wonder if it would survive longer in the most distorted current setting or here...


EDIT EDIT: Well, it just shut down.. I thought it was an empty battery, but no. It's dead.. But it was fun, while it lasted!

Attempt 2: One faint blink, and it was over.. This was a diode, that tested just as high as the first one, during the quick efficiency test...

All i have left now is one very weak diode.. This is the second "worst" diode from my batch.... Not that the simple efficiency test really means much..


Ok, this one survived the first power-up (@251mA), and measured as high as my 6x at 197mA - 189mW. Let's see how long this one lasts..


EDIT: But the wavelength! This is the highest wavelength i have ever seen!

I have some yellow filters, one blocks most 405nm light, and allows comparing wavelengths. The other blocks ALL of it, and leaves only the spontaneous emissions. But this one actually passes through the second filter, and makes a blue spot!

I would almost like to lower the current, to make it last. But it's SO PRETTY!


EDIT 2: Ok, note to self.. 250mA is too much.. :

I double-checked the diodes and even triple-checked one. Most seem to shift mode and have a knee or kink where that happens. For convenience, I'll call the straight spaces in between Tiers (the name just stuck :-? ;D)

This is the one from before. It's the most powerful one, which I previously deemed the weakest. Probably because this is a 4-tier diode, the tiers are a little shorter. IgorT, it still lives!


This one is the one I triple-checked. After I thought I was done increasing it's power, it did even more. This diode exhibits a REAL mode hop, very sharp and even has a hysteresis, where it probably oscillates on the slowness of the temperature (at a very specific current 193.7mA ±0.1mA)


This diode could possibly be a 4-tier also, I didn't dare to go further at the moment because it was the very first one I pushed. I'll revisit this one for sure.


This 3-tier one CODded. Like IgorT says, they'll always pull a new trick if you think you know what you're doing. Pay attention to the hollow curvature of the third tier.


This one oscillated slightly between the first and second tier. It was practically invisible to the eye but the meter kept jumping between values, so I just recorded the max and min value I saw at every mA step. It can probably go higher, but I didn't feel like taking the risk (which increases a lot after 200mW because I have to set the meter to the high range which is 10 times less accurate). Ther's also a small yet sharp mode hop.


Ditto, except here it's between every tier.


This one showed the same negative curvature the dead one did, and I could prevent it from dying. I only wish I did this one before the other one because every tier is shorter :'(.

IgorT said:

Man, i never had to set the range on my meter to high for a blu ray before! ;D

Click to expand...

SAME HERE ;D Now I had to do it three times, and one time i just stopped at 200 because I felt like it.


IgorT, I'm sorry your diodes died. If you read back you see that I found no real correlation between efficiency, other things and the max power, and now that the weak diodes in my list were actually strong, and to make it even worse, are multi-tier, the correlation is non-existent.

Man, I still can't believe it. I bought 10 more diodes because my 190mW from the first four GB diodes died (not from overpowering, which makes it worse) and the next 12 GB diodes were weaker and now I've learned something and I have MULTIPLE diodes STRONGER than a LONG OPEN CAN!!

...and to think that I sold some of my weak diodes at a low price, some even underpowered because the new owners were scared.

I'll continue testing tomorrow. I need a beer! I've been testing diodes for 10 hours straight and it's 2 in the morning.

[edit]changed and added some words here and there.

I'm also thinking about powering them with a two-stage driver. I noticed that when cold, it outputs a lot more. So I think i'll build a driver that switches on at something like 260mW, and after one minute, it switches to the maximum current. This will prevent a photon shock on the facets.

I'm not certain that the switch on is really as bad as some think. Remember these things are actually rated HIGHER when pulsed than they are when run CW. They are rated twice as high when run on nanosecond-width pulses than on CW. I have trouble believed that the power sources are ramping faster than a frequency generator making it up to 100+mW in a few nanoseconds.

While testing, I slowly increase the current, and I got to the point that I measured 240mW at 246mA. When I took the diode from my stash (cold) and put 246mA on it, the reading was 265mW+ and dropping. This means that, say, if 250mW was it's point of death, it would have died from 246mA while it was OK during testing.

pullbangdead I tried to rep you "+2"but it backfired and now you're repped "".

You're trying to say, that it could survive the same current when hot, but not when cold, because the optical flux would be higher, and it would melt the mirrors... Right?

yes, although I also account the possibility that it degrades at an increased rate when the flux is higher and haven't told that before.